Content uploaded by Geoffrey Bird
Author content
All content in this area was uploaded by Geoffrey Bird on Jul 26, 2016
Content may be subject to copyright.
1
Interaction takes two: Typical adults exhibit mind-blindness towards those
with Autism Spectrum Disorder
Rosanna Edey1, Jennifer Cook2, Rebecca Brewer3, Mark H Johnson1, Geoffrey Bird3,4 and
Clare Press1*
1 Department of Psychological Sciences, Birkbeck, University of London
2 School of Psychology, University of Birmingham
3 MRC Social, Genetic, and Developmental Psychiatry Centre. Institute of Psychiatry,
Psychology, & Neuroscience, King’s College London
4 Institute of Cognitive Neuroscience, University College London
* Corresponding author
Address: Department of Psychological Sciences, Birkbeck, University of London, Malet
Street, London, WC1E 7HX
Email: c.press@bbk.ac.uk
Running title: Mind-blindness towards those with autism
Word count: 4984
2
Abstract
Recent work suggests that we are better at interpreting the movements of others who move
like us, and that individuals with Autism Spectrum Disorder (ASD) move in a quantifiably
different way from typical individuals. Therefore, ‘social impairments’ exhibited by
individuals with ASD may, at least in part, represent a failure by typical individuals to infer
the correct mental states from the movements of those with ASD. To examine this possibility,
individuals with ASD and typical adults manually directed two triangles to generate
animations depicting mental state interactions. Kinematic analysis of the generated
animations demonstrated that the participants with ASD moved atypically, specifically with
increased jerk compared to the typical participants. In confirmation of our primary
hypothesis, typical individuals were better able to identify the mental state portrayed in the
animations produced by typical, relative to autistic individuals. The participants with ASD
did not show this ‘same group’ advantage, demonstrating comparable performance for the
two sets of animations. These findings have significant implications for clinical assessment
and intervention in ASD, and potentially other populations with atypical movement.
Keywords: Autism Spectrum Disorder; expertise; action perception; theory of mind
Abbreviations: M=Mean; SEM=Standard error of the mean; AQ=Autism-Spectrum
Quotient; FSIQ=Full scale intelligence quotient; ADOS-2=Autism Diagnostic Observation
Schedule-2; CI=Confidence Interval
3
Summary:
Much research has suggested that people with autism struggle to read the mental states of
others (without autism), but previous work has not investigated how well typical individuals
recognize autistic mental states. Our novel design examines understanding of individuals both
with and without autism by individuals with and without such a diagnosis. Our study suggests
that communicative problems exhibited by individuals with autism may, in part, reflect a
failure by typical individuals to infer the mental states of those with autism.
4
Introduction
Medical professionals, such as general practitioners and mental health specialists, are
frequently required to judge the emotional and mental states of their patients. They will likely
make these judgments on the basis of several cues, including the patient’s verbal report, facial
expressions, postures, and importantly, the way that the patient moves. The kinematics of our
movements provide crucial information about our underlying mental and affective states. For
example, higher velocity movements reflect greater confidence (Fleming, Weil, Nagy, Dolan,
& Rees, 2010), while fast and accelerated movements often reflect anger (Roether, Omlor,
Christensen & Giese, 2009). Thus, in the same way that perception of a smile prompts the
automatic attribution of happiness, perception of fast and accelerated movements, for
example, prompt the attribution of anger (e.g. Atkinson, Tunstall, & Dittrich, 2007).
Association of specific movement cues with specific mental and affective states provides a
rapid route for the attribution of mental and affective states to others, helping clinicians to
detect pain and distress, as well as friends to offer comfort in times of need, and juries and
judges to distinguish deception from sincerity. It has therefore been proposed that perception
of such movement cues is the initial step required for a whole host of socio-cognitive
processes, including theory of mind, which enable rapid and appropriate responses to others
(Klin, Jones, Schultz, & Volkmar, 2003).
However, given that mental states are hidden, perfectly accurate inferences are unlikely to be
made. One variable that is likely to determine the degree of accuracy is whether we move in a
similar way to the observed party. Our experiences with the world tune our perceptual
systems (Blakemore & Cooper, 1970; Sangrigoli, Pallier, Argenti, Ventureyra, & De
Schonen, 2005) and we have extensive experience with our own actions as we learn to
control them (Rochat, 1998; Van der Meer, Van der Weel, & Lee, 1995; White, Castle, &
5
Held, 1964). This experience may tune our perceptual models of action according to how we
move, both through direct visual tuning from self-observation and motor contributions to
perception (Gerson, Bekkering, & Hunnius, 2014; Hunnius & Bekkering, 2014). Therefore,
the way in which we execute actions is likely to have dramatic implications for our
understanding of others, such that we may be better placed both to interpret and interact with
others who move more like us.
We have recently observed that those with Autism Spectrum Disorder (hereafter ‘autism’)
move differently – specifically, with greater jerk, acceleration and velocity – relative to
typical individuals (Cook, Blakemore, & Press, 2013). This finding raises the possibility that
typical individuals experience difficulty interpreting the actions of individuals with autism, in
the same way that those with autism have problems interpreting (e.g. Nackaerts et al., 2012)
and predicting (von der Lühe et al., 2016) typical movements. If action models tuned by
developmental experience of one’s own movements support the recognition of others’ mental
and affective states, then typical individuals, with typical action models, should be impaired
when using atypical movement cues to interpret the mental and affective states of individuals
with autism. Under this hypothesis, interaction difficulties between typical and autistic
1
individuals may be attributable to both parties; typical individuals may make less accurate
mental state attributions about individuals with autism (demonstrate ‘mind-blindness’) in the
same way that individuals with autism appear to display mind-blindness towards typical
individuals. Importantly, if individuals with autism move similarly to each other but
dissimilarly to typical individuals, this hypothesis may help to explain why high-functioning
1
The term ‘autistic people’ is the preferred language of many individuals on the spectrum (see Sinclair, 2013).
In this paper, we use this term as well as person-first language (such as ‘individuals with autism’) to respect the
wishes of all individuals on the spectrum.
6
individuals with autism describe social interactions with other autistic individuals as less
effortful and more efficient than interactions with non-autistic people (Schilbach, 2016).
Much research has investigated how well those with autism are able to understand typical
individuals, but surprisingly, how well typical individuals understand those with autism has
received little attention. Mental state attribution difficulties in typical individuals would not
simply be of academic interest. Medical, legal and educational professionals frequently make
judgments about the intentions and affective states of the individuals with whom they work,
which influence diagnosis, sentencing and intervention planning. If typical individuals in
these roles are required to judge the mental and affective states of autistic individuals without
appropriate models of their movements, errors are likely to be made with significant impact
on the wellbeing of autistic individuals. It is therefore crucial to consider whether typical
individuals have difficulties interpreting the actions of those with autism.
The present study investigated whether typical individuals are impaired at interpreting the
movements produced by autistic, relative to typical, individuals, as well as investigating the
performance of individuals with autism when observing both groups’ movements. We used
the Frith-Happé animation task, as it has been used widely in autism research to assess mental
state attribution (Abell, Happé, & Frith, 2000; Heider & Simmel, 1944). The standard version
of this task presents animations of two triangles moving on a computer monitor, designed by
a non-autistic graphic artist to depict either mental states or random inanimate movement.
Typical children and adults spontaneously attribute appropriate mental states to the former
animation types more readily than those with autism (e.g. Abell, et al., 2000; Castelli, Frith,
Happé, & Frith, 2002). For the present experiment, participants with and without a diagnosis
of autism were asked to direct hand-held triangular magnets on a table top to represent the
7
same mental state verbs as used in the standard paradigm (coaxing, mocking, seducing and
surprising). In a perceptual task several months later, participants watched the animations and
rated the extent to which they depicted each of these target mental states (note that
participants never observed their own generated animations). We compared the accuracy of
ratings in the perception task when typical and autistic participants observed both typical and
autistic animations, to investigate the novel question of whether typical individuals exhibit
specific difficulties when interpreting the movements of those with autism.
Materials and Methods
Participants
Twenty-five typical adults and 23 adults with autism were recruited from the local research
volunteer database for the perceptual task (see Supplementary Methods for information
relating to those participants who generated the animations). An opportunity sample was used
– we contacted all those registered on the database and tested all who volunteered. An
independent clinician diagnosed participants in the autism group according to DSM-IV
criteria (American Psychiatric Association, 2000), and the ADOS-2 (Lord et al., 2012) was
administered to confirm participants’ current level of functioning. Three participants did not
meet all ADOS-2 criteria for a classification of autism spectrum disorder. These participants
were indistinguishable from the other participants on all measures and therefore data from all
autistic participants are reported, but the same pattern of results was found when excluding
these three participants. Two participants (one participant from each group) were excluded
from analysis as their verbal responses suggested that they did not understand the task
demands. The remaining participants were matched on FSIQ (t(44)=.05, p=.964), age (autism
19-69 years, typical 23-74 years, t(44)=.63, p=.531) and gender (Fisher’s exact test, p=.702),
and as expected, the groups differed significantly in AQ scores (t(44)=5.98, p<.001; see
8
Table 1). All participants gave informed consent, and procedures received local ethical
approval. All data were collected in accordance with the guidelines laid out in the 1964
Declaration of Helsinki.
Animation Generation and Kinematic Analysis
A white table with a black enclosure was used as the background for the animations (see
Figure 1). A large red and a small blue triangle made from colored card were attached to two
magnets; the opposite pole of the magnet was attached below the table to enable manual
operation of each triangle. Participants were given time to practice operating the magnets to
maneuver the triangles, and had two practice trials (see Figure 1A; note that they could see
their action effects in the form of the triangle movements). A video camera (Panasonic SDR-
S50) was positioned directly above the table to film the participants’ animations at a rate of
25 frames/second.
The four target mental state words (coaxing, mocking, seducing and surprising – the same as
used in Abell et al., 2000) were presented to participants in a random order. On each trial the
participants were asked: ‘How will you represent (coaxing, mocking, seducing, surprising)
with the two triangles?’ Participants were instructed that their animations should last ~30
seconds, and given one minute to think before providing a verbal response of how they would
animate the mental state word. Participants were directed to the dictionary definition if they
were unsure of the word’s meaning, and no further guidance was given. A follow-up study
demonstrated that independent typical participants could understand the descriptions given by
autistic and typical participants equally well (see Supplementary Methods). Following the
verbal response the participants performed their animation. Each animation was edited to
make size and colors consistent using Adobe After Effects (see Figure 1B, and
9
Supplementary Video 1 for an example video). Any items that appeared on the screen, other
than the triangles, were edited out (e.g. participants’ heads).
The animations were analyzed using MATLAB to extract the kinematics (jerk, acceleration
and velocity) of both triangles. First, every pixel (720x576 pixels) within each frame was
coded for the presence of red or blue and saved as a color-coded frame set. The following
analysis was run twice to code the location of the red and blue triangle separately.
The color-coded frame sets were scanned to locate the four most extreme points of the
triangles (top, bottom, left and right). Two of these points related to the same corner of the
triangle, and therefore one was removed. From the remaining three points, the ‘nose’ of the
triangle was located by identifying the adjoining point at the end of the two longest sides. The
‘tail’ was the mid-point between the other two points. This procedure resulted in location
markers for the nose and tail of the red and blue triangle on each frame. The change in x and
y position was tracked between each frame by first order differentiation of the position
vectors. These vectors were low pass Butterworth filtered at 2Hz to remove noise associated
with the imperfect localization of the x and y values on some frames due to occasional poor
contrast. The velocity was calculated as the square root of the summed squared x and y
displacement per frame. The velocity values were then low pass Butterworth filtered at 3Hz
to remove noise due to the imperfect measure of displacement of the nose/tail owing to
occasional flickers in the animations due to extreme exposure in some of the frames. It was
verified manually that filters were optimal in both preserving the movement information and
removing noise. Visual inspection determined that frames with velocity values below 1
pixel/frame were periods of no movement, and were removed from further calculations.
Acceleration was calculated as the absolute first order differential of the velocity vectors, and
10
jerk was calculated as the absolute second order differential. The mean velocity, acceleration
and jerk was calculated from these absolute values, and transformed from pixels/frame to
mm/second.
Animations were selected for the perceptual task based on two criteria. First, we selected two
animations above and two below the mean jerk value for that group (e.g. autism) and target
mental state (e.g. mocking), all within one standard deviation of the mean, and two
animations greater than one standard deviation from the mean (one above and one below the
mean). Second, we included a maximum of three animations created by any given participant.
These criteria were employed to match the jerk of the selected stimulus set to the distribution
of the full sample of generated animations (see Figure 2A). Further details of the animation
selection are included in Supplementary Methods.
Procedure
The experiment was run via MATLAB on a 24” computer screen. An initial practice trial
familiarized participants with the procedure and task requirements. On each trial, after
viewing the animation, participants were asked to describe verbally what happened in the
animation, to ensure that they had attended to the whole animation. Following the verbal
response, they were asked to rate on a visual analogue scale (ranging from 1 [not at all] to 50
[a lot]) the extent to which the animation depicted the four target mental state words: coaxing,
mocking, seducing, and surprising. Participants were able to rate the four target mental state
words in any order and could change their responses at any point before pressing a key to
begin the next trial. No feedback was given. Six autistic and six typical animations of each of
the four mental state words were presented resulting in 48 animations. Participants who took
11
part in both the animation generation and the perceptual task were not shown their own
animations (see Supplementary Methods).
For each trial, participants’ ratings were scored for accuracy. Accuracy was calculated by
averaging the rating scores of the three ‘incorrect’ scales (e.g. the mean rating of mocking,
seducing and surprising when viewing a coaxing animation), and subtracting this number
from the ‘correct’ scale rating. A score above 0 therefore indicates that the participants rated
the intended mental state verb more highly than the other verbs, with higher scores indicating
a greater match to the intended representation.
Results
The analyses below are reported collapsed across the four mental states (coaxing, mocking,
seducing, surprising). Analyses including mental state as a within-participants factor are
reported in Supplementary Results.
Generated Animation Kinematic Analysis
To verify that the animations generated by the two groups differed in their kinematics, an
independent-sample t-test compared each kinematic measure extracted (jerk, acceleration and
velocity) between the two generator groups (autism or typical; see Figure 2A). Despite no
difference in acceleration (t(29)=1.27, p=.213, 95% CI [-.85, 3.69], d=.46) or velocity values
(t(29)=.17, p=.870, 95% CI [-16.82, 14.30], d=.06), there was a difference with respect to
jerk (t(29)=2.21, p=.035, 95% CI [.08, 2.17], d=.79). Figure 2A demonstrates that this main
effect was driven by increased jerk in the videos produced by individuals with autism,
relative to the typical adults, thus replicating our previous demonstration of increased jerk in
movements produced by individuals with autism (Cook et al., 2013).
12
Perceptual Task
The perceptual task data were analyzed using a mixed 2x2 ANOVA with perceiver group
(autism or typical observers) as a between-participants factor, and generator group (autism or
typical generator) as a within-participants factor.
The main effect of perceiver group was not significant (F(1,44)=2.10, p=.155, 95% CI [-.53,
3.24], η2=.045). There was a main effect of generator group with better rating accuracy when
participants observed the typical (M=7.14, SEM=.61) compared to the autism animations
(M=5.45, SEM=.48; F(1,44)=8.15, p=.007, 95% CI [.49, 2.84], η2=.156). Most importantly
this effect was qualified by an interaction with perceiver group (F(1,44)=4.14, p=.048,
η2=.086). Simple effects analysis showed the typical perceivers were significantly better at
rating the typical animations than the autism animations (p=.001, 95% CI [1.23, 4.48],
d=.66), but the autism perceivers performed equally for the autism and typical animations
(p=.563, 95% CI [-1.22, 2.18], d=.14; Figure 2B).
Discussion
The present experiment required individuals with autism and matched typical adults to
generate animations depicting mental states. The animations generated by the movements of
the participants with autism were found to have greater jerk than their typical counterparts. In
a subsequent perceptual task, the typical participants demonstrated an enhanced ability to
attribute the intended mental state to the animations produced by other typical participants,
relative to those produced by participants with autism. In contrast, individuals with autism
exhibited no difference in assigning the intended mental state to animations produced by
autistic and typical individuals.
13
Differences in movement kinematics between the groups replicate previous findings of
increased jerk when adults with autism make simple horizontal arm movements (Cook et al.,
2013). The current findings extend this work by showing that these atypical kinematics are
evident when those with autism produce complex, object-based actions. A lack of typical
kinematics might be a consequence of peripheral factors such as abnormal muscle tone
(Maurer & Damasio, 1982), or central nervous system factors, such as poor anticipation of
the subsequent part of a motor sequence (Cattaneo et al., 2007). Interestingly, Cook et al.
(2013) also found group differences in velocity and acceleration which were dependent upon
the phase of the movement (turning point or midpoint; see also Forti et al., 2011; Glazebrook,
Elliot, & Lyons, 2006). The phase effects are likely greater for larger magnitude movements
(e.g. full length arm movements), therefore their absence in the present study may be due to
the relatively small magnitude of movements (hand and wrist gestures).
Enhanced perception of typical, relative to autistic actions in the typical perceiver group is
consistent with previous work that suggests we interpret others’ actions according to models
built through experience with our own actions (e.g. Sebanz & Shiffrar, 2009). Typical adults
are better at predicting movement outcomes that comply with typical kinematic trajectories
(Kandel, Orliaguet, & Viviani, 2000), and those with more experience of performing a
particular action are better able to predict the outcome of that action when it is observed
(Diersch et al., 2012). The present study adds to these findings by showing that we are better
at making higher-level mental state inferences (e.g., coaxing) from actions that look like our
own
2
. Future work could compare the extent to which movement kinematics determine our
2
It is important to note that all mental state information must have been derived from movement kinematics in
the current study because there were no facial expressions, postures or language in the animations. However,
despite the fact that jerk is an important cue concerning mental states (e.g. Pollick, Paterson, Bruderlin &
14
ability to make lower- and higher-level inferences from observed actions, perhaps contrasting
these animations against representations of simpler animate concepts (e.g., following).
Our findings have important implications for interaction difficulties between autistic and non-
autistic individuals. Successful and fluid interactions depend on accurate anticipation and
prediction of others’ movements, allowing us to attribute affective states, intentions and goals
to our interaction partner so that we can adjust our behaviour accordingly (Behrends, Müller,
& Dziobek, 2012). Therefore, if typical action models are not effective for interpreting the
movements of individuals with autism, then the ‘social impairments’ exhibited by individuals
with autism may, at least in part, be a product of a failure by typical individuals to infer
correctly the affective states, desires and intentions of autistic individuals. Inappropriate
and/or inconsistent feedback from interaction partners in response to misunderstood actions
may also obstruct learning about typical social interaction in those with autism, interfering
further with social and communication skill development. Crucially, incorrect interpretation
of the movements of individuals with autism by typical individuals could also have important
implications for clinical diagnosis. Autism is diagnosed following observation-based
behavioral assessments of social functioning by a qualified clinician. On the basis of the
current findings we might predict that social expressions by individuals with autism are
harder to decode. Therefore, these individuals may be assessed as lacking social
understanding or expression by non-autistic clinicians, whereas in reality they simply have
models incompatible with those of the assessor. Similarly, other assessments which evaluate
an individuals’ internal states may fail to do so correctly when performed with autistic
individuals, impacting on sentencing, medical support and treatments. Further study of how
Sanford, 2001), other kinematic cues (such as the correlation between the movements) may also have
contributed to inferences.
15
typical individuals could learn to interpret atypical expressions could improve intervention
strategies and also reduce frustration, social anxiety and mood problems in individuals with
autism (Simonoff et al., 2012), who may express their internal states but be frequently
misunderstood.
Despite performing the task at a reasonable level, with comparable performance to the typical
group when observing ‘other’ group animations, the autism group did not show the same
benefit from observing ‘same’ group animations – performance when observing autistic
animations was comparable in the two observer groups. One potential explanation for the
lack of ‘same’ group benefit relates to the trend towards increased variability in the
kinematics of the autistic relative to typical generators. For example, there was a trend for the
four animation-specific jerk values to differ more in each individual in the autism group
(between-group t-test on the standard deviations of the four jerk values; t(29) =1.94, p=.062).
Increased variability in autistic actions may mean that a given autistic participant’s action
models are a poorer fit to those of other autistic individuals (note that autistic facial
expressions have also been shown to have more idiosyncratic qualities, which may similarly
impair expression understanding; Brewer et al., 2016; Macdonald et al., 1989). Alternatively,
autistic action models may be tuned to both atypical (autistic) and typical movements.
Throughout their lives most individuals with autism have much experience observing and
interacting with typical others, who move typically, as well as experience with their own
atypical movements. In contrast, typical individuals are likely to have experience of typical
movements only. These hypotheses could be elucidated by studying perception of autistic
movements in individuals without autism but who have increased experience with this group.
16
The present findings have important implications for other clinical disorders that present with
motor abnormalities. For example, other clinical populations characterized by developmental
(e.g. Tourette Syndrome or cerebral palsy) or neurodegenerative (e.g. Huntington’s Disease
or cerebellar ataxia) motor atypicalities also show atypical intention attribution (Caillies,
Hody, & Calmus, 2012; Eddy & Cavanna, 2015; Eddy & Rickards, 2015; Garrard, Martin,
Giunti, & Cipolotti, 2008). Furthermore, our results may help to explain why children with
Attention Deficit Hyperactivity Disorder (ADHD) and motor dysfunctions are rated as having
more social interaction difficulties than children with ADHD only (Tervo, Azuma, Fogas, &
Fiechtner, 2007). Intentions and emotions expressed by these children with ADHD may be
misunderstood by typical peers, possibly resulting in increased frustration and inadequate
social interactions (Nijmeijer et al., 2008). Even children and adults with motor dysfunction,
without any clinical diagnosis, are more likely to experience less favorable social interactions
(Bejerot & Humble, 2013). Our findings are therefore likely to apply to a number of
conditions characterized by atypical movements, and suggest that interaction difficulties
between typical and atypical individuals may not be attributable solely to the individual who
has received a clinical diagnosis.
17
Acknowledgments
We are grateful to Meredith Leston and Lia Antico for help with testing, and Alex
Chamberlain for assistance in video editing.
18
References
Abell, F., Happé, F., & Frith, U. (2000). Do triangles play tricks? Attribution of mental states
to animated shapes in normal and abnormal development. Cognitive Development, 15(1), 1–
16. http://doi.org/10.1016/S0885-2014(00)00014-9
American Psychiatric Association. (2000). Diagnostic and statistical manual of mental
disorders: DSM-4 (4th ed). Washington, D.C: American Psychiatric Association.
Atkinson, A. P., Tunstall, M. L., & Dittrich, W. H. (2007). Evidence for distinct contributions
of form and motion information to the recognition of emotions from body gestures.
Cognition, 104(1), 59–72. http://doi.org/10.1016/j.cognition.2006.05.005
Behrends, A., Müller, S., & Dziobek, I. (2012). Moving in and out of synchrony: A concept
for a new intervention fostering empathy through interactional movement and dance. The
Arts in Psychotherapy, 39(2), 107–116. http://doi.org/10.1016/j.aip.2012.02.003
Bejerot, S., & Humble, M. B. (2013). Childhood clumsiness and peer victimization: a case–
control study of psychiatric patients. BMC Psychiatry, 13(1), 68. http://doi.org/10.1186/1471-
244X-13-68
Blakemore, C., & Cooper, G. F. (1970). Development of the brain depends on the visual
environment. Nature, 228(5270), 477–478. http://doi.org/10.1038/228477a0
Brewer, R., Biotti, F., Catmur, C., Press, C., Happé, F., Cook, R., & Bird, G. (2016). Can
neurotypical individuals read autistic facial expressions? Atypical production of emotional
facial expressions in Autism Spectrum Disorders. Autism Research, 9(2), 262–271.
http://doi.org/10.1002/aur.1508
Caillies, S., Hody, A., & Calmus, A. (2012). Theory of mind and irony comprehension in
children with cerebral palsy. Research in Developmental Disabilities, 33(5), 1380–1388.
http://doi.org/10.1016/j.ridd.2012.03.012
Castelli, F., Frith, C., Happé, F., & Frith, U. (2002). Autism, Asperger syndrome and brain
mechanisms for the attribution of mental states to animated shapes. Brain, 125(8), 1839–
1849. http://doi.org/10.1093/brain/awf189
Cattaneo, L., Fabbri-Destro, M., Boria, S., Pieraccini, C., Monti, A., Cossu, G., & Rizzolatti,
G. (2007). Impairment of actions chains in autism and its possible role in intention
understanding. Proceedings of the National Academy of Sciences, 104(45), 17825–17830.
http://doi.org/10.1073/pnas.0706273104
Cook, J. L., Blakemore, S.-J., & Press, C. (2013). Atypical basic movement kinematics in
autism spectrum conditions. Brain, 136(9), 2816–2824. http://doi.org/10.1093/brain/awt208
Diersch, N., Cross, E. S., Stadler, W., Schütz-Bosbach, S., & Rieger, M., (2012)
Representing others’ actions: The role of expertise in the aging mind. Psychological
Research, 76(4), 525–41. 10.1007/s00426-011-0404-x
19
Eddy, C. M., & Cavanna, A. E. (2015). Triangles, tricks and tics: Hyper-mentalizing in
response to animated shapes in Tourette syndrome. Cortex, 71, 68–75.
http://doi.org/10.1016/j.cortex.2015.06.003
Eddy, C. M., & Rickards, H. E. (2015). Interaction without intent: The shape of the social
world in Huntington’s disease. Social Cognitive and Affective Neuroscience, 10(9), 1228–
1235. http://doi.org/10.1093/scan/nsv012
Fleming, S. M., Weil, R.S., Nagy, Z., Dolan, R. J., & Rees, G. (2010). Relating introspective
accuracy to individual differences in brain structure. Science, 329((5998), 1541-1543.
10.1126/science.1191883
Forti, S., Valli, A., Perego, P., Nobile, M., Crippa, A., & Molteni, M. (2011). Motor planning
and control in autism. A kinematic analysis of preschool children. Research in Autism
Spectrum Disorders, 5(2), 834–842. http://doi.org/10.1016/j.rasd.2010.09.013
Garrard, P., Martin, N. H., Giunti, P., & Cipolotti, L. (2008). Cognitive and social cognitive
functioning in spinocerebellar ataxia: A preliminary characterization. Journal of Neurology,
255(3), 398–405. http://doi.org/10.1007/s00415-008-0680-6
Gerson, S. A., Bekkering, H., & Hunnius, S. (2014). Short-term motor training, but not
observational training, alters neurocognitive mechanisms of action processing in infancy.
Journal of Cognitive Neuroscience, 1–9. http://doi.org/10.1162/jocn_a_00774
Glazebrook, C. M., Elliott, D., & Lyons, J. (2006). A kinematic analysis of how young adults
with and without autism plan and control goal-directed movements. Motor Control, 10(3),
244.
Hunnius, S., & Bekkering, H. (2014). What are you doing? How active and observational
experience shape infants’ action understanding. Philosophical Transactions of the Royal
Society B: Biological Sciences, 369(1644), 20130490–20130490.
http://doi.org/10.1098/rstb.2013.0490
Heider, F., & Simmel, M. (1944). An experimental study of apparent behavior. The American
Journal of Psychology, 57(2), 243. http://doi.org/10.2307/1416950
Kandel, S., Orliaguet, J.-P., & Viviani, P. (2000). Perceptual anticipation in handwriting: The
role of implicit motor competence. Perception & Psychophysics, 62(4), 706–716.
http://doi.org/10.3758/BF03206917
Klin, A., Jones, W., Schultz, R., & Volkmar, F. (2003). The enactive mind, or from actions to
cognition: Lessons from autism. Philosophical Transactions of the Royal Society B:
Biological Sciences, 358(1430), 345–360. http://doi.org/10.1098/rstb.2002.1202
Lord, C., Rutter, M., DiLavore, P., Risi, S., Gotham, K., & Bishop, S. (2012). Autism
Diagnostic Observation Schedule: ADOS-2 (2nd ed.). Torrance, CA: Western Psychological
Services.
Macdonald, H., Rutter, M., Howlin, P., Rios, P., Conteur, A. L., Evered, C., & Folstein, S.
(1989). Recognition and Expression of Emotional Cues by Autistic and Normal Adults.
20
Journal of Child Psychology and Psychiatry, 30(6), 865–877. http://doi.org/10.1111/j.1469-
7610.1989.tb00288.x
Maurer, R. G., & Damasio, A. R. (1982). Childhood autism from the point of view of
behavioral neurology. Journal of Autism and Developmental Disorders, 12(2), 195–205.
http://doi.org/10.1007/BF01531309
Nackaerts, E., Wagemans, J., Helsen, W., Swinnen, S. P., Wenderoth, N., & Alaerts, K.
(2012). Recognizing biological motion and emotions from point-light displays in Autism
Spectrum Disorders. PLoS ONE, 7(9), e44473. http://doi.org/10.1371/journal.pone.0044473
Nijmeijer, J. S., Minderaa, R. B., Buitelaar, J. K., Mulligan, A., Hartman, C. A., & Hoekstra,
P. J. (2008). Attention-deficit/hyperactivity disorder and social dysfunctioning. Clinical
Psychology Review, 28(4), 692–708. http://doi.org/10.1016/j.cpr.2007.10.003
Pollick, F. E., Paterson, H, M., Bruderlin, A., & Sanford, A. J. (2001). Perceiving affect from
arm movement. Cognition, 82, B51-B61.
Rochat, P. (1998). Self-perception and action in infancy. Experimental Brain Research,
123(1-2), 102–109. http://doi.org/10.1007/s002210050550
Roether, C. L., Omlor, L., Christensen, A., & Giese, M. A. (2009). Critical features for the
perception of emotion from gait. Journal of Vision, 9(6), 15–15. http://doi.org/10.1167/9.6.15
Sangrigoli, S., Pallier, C., Argenti, A. M., Ventureyra, V. A. G., & De Schonen, S. (2005).
Reversibility of the other-race effect in face recognition during childhood. Psychological
Science, 16(6), 440–444.
Schilbach, L. (2016). Towards a second-person neuropsychiatry. Philosophical Transactions
of the Royal Society B: Biological Sciences, 371(1686), 20150081.
http://doi.org/10.1098/rstb.2015.0081
Sebanz, N., & Shiffrar, M. (2009). Detecting deception in a bluffing body: The role of
expertise. Psychonomic Bulletin & Review, 16(1), 170–175.
http://doi.org/10.3758/PBR.16.1.170
Simonoff, E., Jones, C.R., Pickles A., Happé, F., Baird, G., & Charman, T. (2012). Severe
mood problems in adolescents with autism spectrum disorder. Journal of Child Psychology
and Psychiatry, 53(11), 1157-1166. 10.1111/j.1469-7610.2012.02600.x
Sinclair, J. (2013). Why I dislike ‘‘person first’’ language. Autonomy, the Critical Journal of
Interdisciplinary Autism Studies, 1.2: 1-2.
Tervo, R. C., Azuma, S., Fogas, B., & Fiechtner, H. (2007). Children with ADHD and motor
dysfunction compared with children with ADHD only. Developmental Medicine & Child
Neurology, 44(6), 383–390. http://doi.org/10.1111/j.1469-8749.2002.tb00832.x
Van der Meer, A. L. H., Van der Weel, F. R., & Lee, D. N. (1995). The functional
significance of arm movements in neonate. Science, 267, 693–695.
21
Von der Lühe, T., Manera, V., Barisic, I., Becchio, C., Vogeley, K., & Schilbach, L. (2016).
Interpersonal predictive coding, not action perception, is impaired in autism. Philosophical
Transactions of the Royal Society B: Biological Sciences, 371(20150373), 1–8.
White, B. L., Castle, P, & Held, R. (1964). Observations on the development of visually-
directed reaching. Child Development, 35, 349–364.
22
Table 1: Demographic information for autistic and typical participants in the
perceptual task
**p<.001
FSIQ
mean
(SEM)
Age
Mean
(SEM)
Gender
male
AQ
mean
(SEM)
ADOS
mean
(SEM)
Autism
Perceiver
Group
(n =22)
111.05
(3.62)
38.50
(2.97)
19
32.82
(2.11)**
9.77
(.79)
Typical
Perceiver
Group
(n=24)
111.25
(2.69)
36.08
(2.45)
19
17.29
(1.56)
-
23
Figure 1: A) Participants generated animations by manually directing triangles on a
table top with magnets to represent the target mental state words (coaxing, mocking,
seducing and surprising). B) Edited example stimulus that was displayed to participants
as a ~30 second animation.
24
Figure 2. A) Mean velocity, acceleration and jerk for the autism and typical animations.
B) Mean rating accuracy for the autism and typical groups when rating the autism and
typical generated animations. Error bars represent the standard error of the mean.
*p<.05
25
Interaction takes two: Typical adults exhibit mind-blindness towards those
with Autism Spectrum Disorder – Supplementary Materials
Methods:
Animation Generation - Participants
Fifteen typical adults and 16 adults with Autism Spectrum Disorder (hereafter ‘autism’) were
recruited from the local research volunteer database to generate the animations. The same
recruitment method was used as for the main perceptual task. The participants did not differ
significantly on FSIQ, as measured by the Wechsler Abbreviated Scale of Intelligence
(Wechsler, 1999, t(29)=2.00, p=.054), age (autism 28-68 years, typical 20-57 years,
t(29)=1.69, p=.102) or gender (Fishers exact test, p>.99), but as expected the autism group
had significantly higher scores on the AQ (t(23.34)=6.15, p<.001, see Supplementary Table
1). All participants in the autism group had a clinical diagnosis of autism from a qualified
independent clinician according to DSM-IV criteria, however two participants did not meet
all ADOS-2 criteria for a classification of autism spectrum disorder. These participants were
indistinguishable from the other participants on all measures and therefore data from all
autistic participants are reported, but the same pattern of results was found when excluding
these two participants.
Animation Selection
Thirteen of the autism group and eight of the typical group also participated in the main
perceptual task. There were approximately nine months between the animation generation
and perceptual task. Participants who took part in both animation generation and the main
perceptual task were not shown their own animations. Their animations were substituted with
26
another exhibiting the closest jerk value for that animation (e.g. mocking) within their group
(autism/typical).
Verbal Description Experiment
To verify that both generator groups had sufficient understanding of the mental state concepts
a follow up study was conducted with an independent group of typical participants (N=24),
who were matched for age (23 – 67 years, t(46)=.21, p=.837) and gender (χ2(2)=.12, p=.731)
to the typical participants in the perception task. Participants were presented, in a random
order, with the verbal descriptions given by the generators before producing the animations
used in the perception task. The method was matched closely to that employed in the
perception task. Participants rated out of 50 the extent to which the description represented
each of the four target mental states (coaxing, mocking, seducing and surprising). We
selected only those descriptions where the target verb was not used, and randomly selected
from the remaining set to equate the number of autistic and typical descriptions. This method
resulted in three surprising and mocking descriptions from each group and four coaxing and
seducing descriptions, totaling 28 descriptions.
Results:
Verbal Descriptions
The mean ratings for both the autism and typical descriptions were found to be significantly
different from zero (autism t(23)=15.89, p<.001; typical t(23)=22.65, p<.001), indicating that
both groups understood the concepts. To look for differences between these ratings, we
performed a t-test comparing the ratings, which revealed no effect of generator group
(t(1,23)=1.24, p=.226, 95% CI [-1.01, 4.06], d=.25; autism M=22.38, SEM=1.41; typical:
M=20.85, SEM=.92). The absence of this effect demonstrates that the verbal descriptions
27
provided by each group were comprehended equally, thereby indicating that both generator
groups had equal understanding of the concepts.
Generated Animations: Effects and interactions involving mental state animation
A mixed 2x4 ANOVA was performed for each kinematic measure with group (autism or
typical) as a between-participants factor, and mental state (coaxing, mocking, seducing or
surprising) as a within-participants factor. Of course the main effects of group remained
unchanged with the addition of the extra factor, and therefore are only reported in the Main
Text.
The group effect of jerk did not interact with mental state animation (F(3,87)=.132, p=.941,
η2=.005). However, there was a significant main effect of mental state animation in jerk
values (F(3,87)=8.74, p<.001, η2=.232). Follow up pairwise comparisons revealed the
mocking animations (M=6.52mm/s3, SEM=.33) had higher mean jerk than the coaxing
(M=4.46mm/s3, SEM=.338; p=.003, 95% CI [.55, 3.56], d=.71) and seducing animations
(M=4.57mm/s3, SEM=.34; p=.002, 95% CI [.57, 3.32], d=.73), and the surprising animations
(M=5.57mm/s3, SEM=.32) had higher mean jerk than the coaxing animations (p=.016, 95%
CI [.151, 2.07], d=.60). The lack of interaction between mental state animation and group -
and therefore comparable pattern of kinematics across the four mental state animations (see
Supplementary Figure 1) - suggests that both groups modulated the kinematics features of
their movements to represent different mental states in a similar fashion. This result is
consistent with the suggestion that both groups possessed a similar understanding of the
different mental states (see Verbal Descriptions Experiment).
28
There was no interaction between mental state animation and group (F(3,87)=.33, p=.806,
η2=.011) in acceleration values. However, there was a main effect of animation
(F(3,87)=9.73, p<.001, η2=.251). Follow-up pairwise comparisons revealed the mocking
animations (M=15.36mm/s2, SEM=1.05) had higher mean acceleration than the coaxing
(M=10.71mm/s2, SEM=.83; p=.003, 95% CI [1.31, 7.99], d=.72) and seducing animations
(M=11.06mm/s2, SEM=.69; p=.003, 95% CI [1.20, 7.40], d=.70), the surprising animations
(M=13.94mm/s2, SEM=.74) had higher mean acceleration than the coaxing animations
(p=.001, 95% CI [1.13, 5.33], d=.79), and the surprising animations had higher mean
acceleration than the seducing animations (p=.038, 95% CI [.12, 5.65], d=.53).
There was no interaction between mental state animation and group (F(3,87)=.38, p=.766,
η2=.013) in velocity. As with the jerk and acceleration analyses, there was a main effect of
mental state animation (F(3,87)=9.08, p<.001, η2=.238). Follow-up pairwise comparisons
revealed the mocking animations (M=109.05mm/s, SEM=7.23) had higher mean velocity
than the coaxing (M=80.77mm/s, SEM=4.57; p=.003, 95% CI [7.84, 48.73], d=.71) and
seducing (M=83.62mm/s, SEM=4.53; p=.015, 95% CI [3.61, 47.26], d=.59) animations, the
surprising animations (M=101.29mm/s, SEM=5.12) had higher velocity than the coaxing
animations (p=.001, 95% CI [7.74, 33.31], d=.83), and the surprising animations had higher
mean velocity than the seducing animations (p=.050, 95% CI [.02, 35.33], d=.52).
Perceptual task: Effects and interactions involving mental state animation
Both the autism (M=5.63, SEM=.70, t(21)=8.05, p<.001, 95% CI [4.18, 7.09], d=1.76) and
typical (M=6.96, SEM=.62; t(23)=11.29, p<.001, 95% CI [5.68, 8.23], d=2.30) perceiver
groups’ overall mean accuracy was significantly higher than zero, indicating that both groups
could do the task effectively. Additionally, the mean accuracy for the animations generated
29
by the autism (M=5.48, SEM=.48; t(45)=11.48, p<.001, 95% CI [4.51, 6.44], d=1.28) and
typical (M=7.19, SEM=.63; t(45)=11.45, p<.001, 95% CI [5.93, 8.46], d=2.56) participants
was higher than zero, demonstrating significant signal present in the animations generated by
both groups.
A mixed 2x2x4 ANOVA was performed with perceiver group (autism or typical observer) as
a between-participants factor and generator group (autism or typical generator) and mental
state (coaxing, mocking, seducing or surprising) as within-participants factors. Of course the
main effects of perceiver group and generator group, as well as their interaction, remained
unchanged with the addition of the extra factor, and therefore are only reported in the Main
Text.
There was a main effect of mental state animation (F(3,132)=43.13, p<.001, η2=.496), which
did not interact with perceiver group (F(3,132)=1.70, p=.171, η2=.037) but did interact with
generator group (F(3,132)=7.10, p<.001, η2=.139). This interaction was driven by better
performance on the typical (M=6.16, SEM=.61) relative to autism surprising animations
(M=-1.16, SEM=.48; p<.001, 95% CI [4.43, 10.21], d=.73), without a generator group
difference for the other mental state animations.
30
Supplementary Table 2: Demographic information for those participants who
generated animations and those who took part in the verbal descriptions experiment
Supplementary Figure 1: Mean jerk for the four mental states for the typical and autistic
animations. It can be seen that the jerk effect did not interact with mental state, such that both
groups modulated their kinematics comparably according to the different mental states.
FSIQ
mean
(SEM)
Age
Mean
(SEM)
Gender
male
AQ
mean
(SEM)
ADOS
mean
(SEM)
Autism
Generator
Group
(n =16)
116.00
(3.04)
42.06
(2.96)
14
30.44
(2.24)**
9.63
(.96)
Typical
Generator
Group
(n =15)
105.53
(4.32)
34.80
(3.12)
13
14.67
(1.25)
-
Typical
Verbal
Experiment
Group
(n=24)
-
35.46
(2.98)
18
-
-
A preview of this full-text is provided by American Psychological Association.
Content available from Journal of Abnormal Psychology
This content is subject to copyright. Terms and conditions apply.
